Precision of Survey Estimates Derived from the Medical Expenditure Panel Survey (MEPS)

The sample design of the Medical Expenditure Panel Survey Household Component (MEPS-HC) is characterized by a complex multi-stage area probability design that includes disproportionate sampling of specified policy relevant population groups. As a consequence of departures from simple random sampling assumptions, the variances of survey estimates derived from the MEPS will generally exhibit design effects that are substantially greater than unity. A summary of design effect variations for 19961998 MEPS estimates has been previously reported (Yu, W., 2002). Based on data from the 1999 and 2000 MEPS-HC, this paper will evaluate and contrast the design effects achieved for national estimates of health care utilization, expenditures, and sources of payment; the level of design effect variation in related survey estimates; and design effect variation by alternative population subgroups and by different geographic regions of the nation. This analysis will also include an evaluation of design effects achieved for health care estimates associated with individuals that incur high levels of medical expenditures. The results will help improve the sample design specifications for the selection of future new sample panels of households for the annual MEPS-HC.